Selection Bias in Online Experimentation

Take a break for a few minutes outside “Cairo’,” one of our kitchens in San Francisco. The coffee is hot, and the view is fabulous. In online experimentation platforms, we choose the experiments with significant successful results to launch to the product. When estimating the aggregated impact of the launched features, we investigate a statistical selection bias in this process and propose a corre

[simomm]

個別施策のABテスト結果による押上分を足しても、全施策をリリースした時の押上にはならないという話。